Comparison of Fourier and wavelet analysis for fatigue assessment during repetitive dynamic exertion

The comparative ability of the Fourier transform (FFT) and discrete wavelet transform (DWT) algorithms in assessing muscle fatigue during sub-maximal repetitive dynamic exertion was investigated in this study. Surface electromyography data recorded from the upper trapezius muscle during forty minutes of repetitive upper extremity exertion performed by 10 male participants were used in the analysis. Multi-model regression analysis was performed to study the trend in the power values of the different frequency bands estimated using the FFT and DWT algorithms. Less variability and higher statistical significance was observed for the power value trend computed using the DWT algorithm compared to the FFT algorithm. The regression models provided a better fit for the power values estimated under more fatigued condition compared to the less fatigued condition. The lower frequency bands of 23–46 Hz and 46–93 Hz exhibited the expected and consistent power trend independent of the algorithm (DWT or FFT) used. For the exertions tested in this study, a cubic or curvilinear model explained the fatigue development process with a higher precision than the linear models.